Russell and Kaplan on Denoting

Abstract

For years now philosophers have established their reputations by offering up new and memorable criticisms of Russell's beleaguered Theory of Descriptions. Strawson' argued, as opposed to Russell, that existential import of definite descriptions should be viewed rather as pragmatic presupposition than as semantic content. Presupposition-failure would thus lead to truth-value gaps rather than straightforward falsity. Donnellan' maintained that Russell's theory at best accounted for descriptions used 'attributively', where referent is strictly denoted, but failed to explicate truly 'referential' use of such terms. Others3 have pointed out that (whether or not Donnellan is right) most descriptions in everyday speech are in fact 'improper'the content is not fully enough specified to actually determine precise referent intended. And Sharvy4 has recently urged that a 'more general theory of descriptions' should account for massas well as count-nouns. But one of most interesting, and radical, criticisms of all is offered by David Kaplan in his remarkable article, What is Russell's Theory of Descriptions?5 Kaplan points out that most distinctive feature of Russell's theory has not attracted notice it deserves to wit, fact that for Russell descriptions do not really denote at all. For Russell a statement, the F is G or G(LxFx) is ultimately analyzed as: (3x) (y) (Fy y x . Gx). The latter, clearly, contains no singular terms except variables,